Published October 17, 2023
| Version v1
Model
Open
Data release of Being confident in confidence scores: calibration in deep learning models for camera trap image sequences
Description
Data release of the submitted journal article : Being confident in confidence scores: calibration in deep learning models for camera trap image sequences
Article is available on bioRxiv : https://www.biorxiv.org/content/10.1101/2023.11.10.566512v2
It includes :
- The five pretrained models
- An inference script : compute_features.py
- Metadata of the 3 test sets
- Cropped data of EcrinsSCJ2 datasets (Alps test set)
- Precomputed features for the 5 models x 3 test sets
- A script to compute ECE and Accuracy values : compute_calibration.py
Files
cropped_data.zip
Files
(3.2 GB)
| Name | Size | Download all |
|---|---|---|
|
md5:6d8630f917bfb04b6639876e9071f050
|
2.4 kB | Download |
|
md5:4695d4e2391ba14ceab95eb4623aedc8
|
4.0 kB | Download |
|
md5:73b6f3bddf2cf5e7a514fb58805d6dc0
|
3.5 kB | Download |
|
md5:9247a76b76f3817f6feacfb72fc17d60
|
350.5 MB | Download |
|
md5:4e1c76f21e4a3b686263f31b869f98c5
|
1.8 GB | Preview Download |
|
md5:53c00879d9a41b8a9cf52a5b0564fb2a
|
89.7 MB | Download |
|
md5:729141ebd7351b71361a5ad4ac980c36
|
1.8 MB | Preview Download |
|
md5:89ef131e1f7574ce31ba50fafbca449f
|
16.7 MB | Preview Download |
|
md5:deed2e0b5e81fc3b079e76767b724c31
|
18.3 MB | Preview Download |
|
md5:fa63a8bc7e096f21b5f4b012c7b4da7a
|
17.1 MB | Download |
|
md5:6e4d4690b4391f21d87b66bf341c6f80
|
43.5 MB | Download |
|
md5:0f089c0e26d567a27ac7a8d982df570e
|
43.5 MB | Download |
|
md5:5cdeb99057c748760f384f76972c5d2c
|
43.5 MB | Download |
|
md5:ad8be8e0510261e8ff46eee4e313a526
|
43.5 MB | Download |
|
md5:f9184c41d47e4bb161db10f323f21ee7
|
43.5 MB | Download |
|
md5:94966e0fddc6ce9471969a087d18921e
|
347.8 MB | Download |
|
md5:3760a25d83aab7537b0078100e90553f
|
343.3 MB | Download |
Additional details
Related works
- Is described by
- 10.1101/2023.11.10.566512 (DOI)